USPTO Patents Application 09230955

This Page Is Inserted by IFW Operations
and is not a part of the Official Record

BEST AVAILABLE IMAGES

Defective images within this document are accurate representations of
the original documents submitted by the applicant.

Defects in the images may include (but are not limited to):

BLACK BORDERS

TEXT CUT OFF AT TOP, BOTTOM OR SIDES
FADED TEXT
ILLEGIBLE TEXT
SKEWED/SLANTED IMAGES
COLORED PHOTOS

BLACK OR VERY BLACK AND WHITE DARK PHOTOS
GRAY SCALE DOCUMENTS

IMAGES ARE BEST AVAILABLE COPY.

As rescanning documents will not correct images,
please do not report the images to the
Image Problem Mailbox.

STIC-I

From:
Sent:
To:
Subject:

to. .

Canella, Karen

Wednesday, May 14, 2003 3:05 PM
STIC-ILL

ill order 09/230,955

Art Unit 1642 Location 8E12(mail)

Telephone Number 308-8362

Application Number 09/230,955

1 . American Journal of Pathology:

1993 Feb, 142(2):403-412
1993, 143(4): 11 50-1 158
1984, 114(3):454-460
1996, 148(3):865-875
1965 Sep, Vol. 44, pp. 280-282

2.
3.
4.

5.
6.

8.
9.
10.
11.

12.
13.
14.

Cancer Research, 1993 May 15, 53(10 suppl):2287-2299

Cancer epidemiology, biomarkers and Prevention, 1996 Jul, 5(7):549-557

Lab Investigation:

1980. 42(1):91-96
1988, 58(2):141-149

Gynecol Oncol, 1 982, 1 3(1 ):58-66

International Journal of Gynecological Pathology:
1985,4(4):300-313
1986, 5(2): 151 -162
1992, 11(1):24-29

^ is

Cancer (Phila), 1989, 63(7): 1337-1 342
Cancer Res, 1990, 50(16):5143-5152

Virchows Arch B Cell Pathol Incl Mol Pathol, 1987, 54 (2):98-110

Acta Histochemica et Cytochemica:
1994, 27(3):251-257
1996, 29(1 ):51 -56

Archives of Gynecology and Obstetrics, 1989, 246(4):233-242

Clin Lab Med, 1995 Sep, 15(3):727-742

Clin Obstet Gynaecol, 1984 Apr, 1 1 (1):5-23

l

HI Differentiation (1986) 31 : 191-205

Is and
inoma
"nono-

factor
arcin-

[1981)
rcino-

ation.

m for
3

An-
cntia-
lature

r-dip-
i that

yonal
ation

ercn-
nt of
pro-
I Res

lduc-
ihibi-

>n of
:rmal

blem

:iiiTs,

1985)
ident
n re-
IRcs

t for
oma

ning
Natl

The
cell

uced
Is in

> for

t

lion
403
duc-
era-
MP.

4ar-
em-

86

Differentiation

£> Springer-Verlag 1986

Cytokeratin expression in squamous metaplasia of the
human uterine cervix

Orith Gigi-Leitner 1 , Benjamin Geiger 1 *, Rivka Levy 13 , and Bernard Czernobilsky 1 ' 2

1 Department of Chemical Immunology, The Weizmann Institute of Science Rehovot, Israel

2 Department of Pathology, Kaplan Hospital, Rehovot, Israel**

3 Department of Pathology, Tel Aviv University, Tel Aviv, Israel

Abstract. The expression of cytokeratin polypeptides in
squamous metaplasia of the human uterine cervix was in-
vestigated by immunocytochemical labeling with polypep-
tide-specific antibodies against cytokeratins. Immunofluor-
escence microscopic examination of cervical tissues using
various monoclonal antibodies indicated that squamous
cervical metaplasia expresses a unique set of cytokeratin
polypeptides, this being distinctively different from that ex-
pressed by all of the normal epithelial elements of the exo-
and endocervix. The development of metaplastic foci was
accompanied by the expression of cytokeratin polypeptide
no. 13, which is commonly detected in stratified epithelia,
and by a reduction in the level of polypeptide no. 1 8, which
is typical of simple epithelia. The 40-kilodalton cytokeratin
(no. 19) described by Moll etal., which is abundant in the
columnar and reserve cells of the endocervix, was found
throughout the metaplastic lesions. Only in ' well-differen-
tiated' metaplasias did we detect polarity of cytokeratin
expression reminiscent of the staining patterns in the exo-
cervix. This was manifested by the exclusive labeling of
the basal cell layer(s) with antibodies K B 8.37 and K M 4.62,
which stain the basal cells of the exocervix. Furthermore,
a comparison of cervical metaplasia with squamous areas
occurring within endometrial adenocarcinomas pointed to
a close similarity in the cytokeratin expression of the two.
We discuss the use of cytokeratins as specific markers of
squamous differentiation, the relationships between squa-
mous metaplasia and cervical neoplasia, and the involve-
ment of reserve cells in the metaplastic process.

Introduction

The process of squamous metaplasia involves the transfor-
mation of differentiated nonsquamous epithelium into
squamous epithelium [31, 46]. This process is often detected
in various human tissues such as the bronchi [3, 49], stom-
ach [9, 32], urinary bladder [25], and salivary glands [11],
and is especially common in the uterine cervix [12, 13].
The human cervix consists of two major areas that are ana-
tomically and histologically distinct: the exocervix and en-
docervix. The former is characteristically composed of non-
keratinizing stratified squamous epithelium, while the latter

* To whom offprint requests should be sent
** Affiliated with the Medical School of The Hebrew University
and Hadassah Hospital, Jerusalem, Israel

contains a simple epithelial monolayer of columnar cells
which line the mucosal surface and invaginate into the
stroma. The sharp squamocolumnar junction detected be-
tween the two areas is normally located at the cervical por-
tio [8, 12].

This boundary area has been found to be the most com-
mon site for the development of squamous metaplasia [6,
12, 43]. This metaplasia is preceded by the outward exten-
sion of the endocervical mucosa into the exocervical por-
tion, a process denoting erosion or ectopy [12]. Subse-
quently, and usually throughout the reproductive period
of the individual, changes may occur in the ectopic endo-
cervical mucosa, leading to various degrees of stratification
of the epithelium and the formation of 'transformation-
zone metaplasia' (TZM). The metaplastic squamous epithe-
lium thus formed is usually less ordered than that of the
neighboring exocervix and can usually be identified by con-
ventional light microscopy [12]. Nevertheless, is often re-
tains an apparent continuity with the exocervical epithelial
layers, and the exact point of transformation is difficult
to determine (see Fig. 1 c).

Less frequent, yet still quite common, is the appearance
of * metaplastic plaques' (MPs) in the endocervical canal,
which are completely disconnected from any normal squa-
mous epithelium. These MPs may exhibit variable dimen-
sions and may differ with regard to the extent of stratifica-
tion. Similar metaplastic changes may be found not only
in the normal mucosa but also in malignancies of glandular
tissues of the female genital tract, including the endocervix,
forming 'adenocarcinomas with squamous differentiation*
[26].

The factors which induce metaplastic transformation
are still poorly understood, yet several possibilities have
been suggested. These include alterations in environmental
conditions, mechanical irritation, chronic inflammation,
changes in pH and in hormonal balance, etc. [12, 43]. An-
other debatable issue concerns the cellular basis of meta-
plastic transformation : is there a direct ingrowth from the
native portio epithelium into the transformation zone ([43];
or even beyond it), or does the process involve cells of
purely endocervical origin? Attempts to identify cellular
precursors for squamous metaplasia exhibiting the latter
mechanism have focused on two endocervical cell types,
i.e., * basal* or 'reserve* cells and columnar cells. As will
be discussed in detail later, the former are commonly scat-
tered in between the columnar cells of the endocervix and
confined to the area close to the basement membrane (for

192 "

further details, see Discussion). For a number of years,
these reserve cells have been accepted as being the source
of cervical squamous metaplasia, a theory originally pro-
posed by Fluhmann, who called this process "prosoplasia"
[14]. Until recently, the major, if not only, means of study-
ing metaplastic transdifferentiation was morphological ob-
servations using light and electron microscopy.

Recently, immunocytochemical techniques employing
cell-type-specific antibodies have been widely used for the
identification of the histogenetic origins as well as the state
of differentiation of cells. Particularly useful in this respect
are specific antibodies reactive with intermediate filament
(IF) subunits (for reviews, see [17, 36, 37, 41, 51]). It has
been extensively documented that there are five major, bio-
chemically and antigenically distinct families of IF subunits
which are expressed in a cell-type-restricted fashion [2, 28,
29]. Among those, the cytokeratin family, which is charac-
teristic of epithelial cells, is further diversified [16, 33, 40]:
about 20 different cytokeratin polypeptides from various
human epithelia have been isolated and biochemically, im-
munochemically, and genetically characterized [18, 23, 33,
40, 45]. It has further been shown that each type of epitheli-
al cell contains a characteristic combination of cytokeratin
polypeptides which may be used to identify that particular
cell type either in the normal state or after malignant trans-
formation (for reviews, see [33, 40, 47]). This approach
has been extensively employed in recent years for the diag-
nosis of anaplastic tumors and the determination of their
histogenetic origins [4, 19, 34, 42].

In the present study, we investigated the expression of
specific cytokeratin polypeptides in different forms of squa-
mous metaplasia of the human cervix. Using both biochem-
ical and immunohistochemical approaches, we showed that
cells undergoing metaplastic changes express a unique com-
bination of cytokeratin polypeptides including the stratifi-
cation-related cytokeratin polypeptide no. 13 (which is oth-
erwise absent from the normal mucosa of the endocervix),
polypeptides nos. 8 and 19, and minute and variable
amounts of polypeptides nos. 18 and 10/11 (numbers ac-
cording to the classification of Moll et al. [33]). These re-
sults suggest that the metaplastic process involves a unique
step of squamous differentiation of an endocervical cell
(probably a reserve cell) which is molecularly distinct from
the process of stratification of the exocervix. The signifi-
cance of these results and their relevance to cervical neopla-
sia are discussed.

Methods

Tissues

The cervical tissues studied were obtained from 31 patients
aged 36-82 years (mean, 54 years) at the Kaplan Hospital,
whose uteri were removed due to leiomyomas and prolapse.
The cervix was opened through the external os within
15 min of hysterectomy, and several sections were obtained
through the exocervix and endocervix, including the squa-
mocolumnar junction in a plane parallel to the long axis
of the cervical canal. The tissues used for immunocytochem-
ical studies were snap frozen in isopentane that had been
precooled in liquid nitrogen, and then stored at -70°C.
For routine histologic examinations, the tissues were fixed
in 4% buffered formaldehyde, embedded in paraffin [1],
and stained with hematoxylin and eosin (HE). In 8 out

of the 31 cases examined, various degrees of cervical squa-
mous metaplasia were observed, either in continuity with
the exocervix or in isolated foci within the cervical canal.

Immunochemical reagents

The murine monoclonal antibodies used included:

1. K G 8.13, a broad-spectrum cytokeratin antibody
which reacts with the cytokeratin filaments present in all
human epithelial cells tested, i.e., both normal and malig-
nant cells. This antibody, raised against bovine muzzle kera-
tin, reacts with a relatively broad range of polypeptides
including cytokeratins nos. 1, 5, 6, 7, 8, and 18, as well
as reacting weakly with cytokeratins nos. 10 and 1 1 [22],

2. K K 8.60, an antibody reactive with human cytokera-
tin polypeptides nos. 10 and 11. As previously suggested,
this antibody might be a specific marker of keratinization
[24].

3. K s 8.12, an antibody that reacts with polypeptides
nos. 13 and 16, which are present in stratified nonkeratiniz-
ing epithelia as well as in squamous carcinomas [24],

4. K B 8.37, an antibody which reacts with IFs of cul-
tured keratinocytes of murine and bovine origin (data not
shown), as well as with cytokeratin filaments in the basal
layer of stratified squamous epithelium (skin, exocervix,
etc. ; see insert in Fig. 2e). This antibody does not react with
simple, pseudostratified or transitional epithelia in humans.
The exact polypeptide specificity of this antibody has not
been defined, since it does not react with electrophoretically
separated polypeptides of the exocervix in Western-blot
analysis. The epitope specifically recognized by this anti-
body may be conformation dependent and thus be irreversi-
bly destroyed by electrophoretic separation. Regardless of
its fine molecular specificity, we used antibody K B 8.37 as
a marker of the basal layer of the squamous epithelium.

5. K M 4.62, a monoclonal antibody prepared against cy-
toskeletal polypeptides of cultured human adenocarcinoma
line MCF-7. This antibody reacts with only one human
polypeptide, i.e., no. 19 [21].

6. K s 18.18, a murine monoclonal antibody which reacts
with human cytokeratin polypeptide no. 18 and stains sim-
ple and pseudostratified epithelia as well as the basal layer
of several noncornifying stratified squamous epithelia
(W.W. Franke, unpublished data). This antibody was
kindly supplied by Prof. W.W. Franke (German Cancer
Research Center, Heidelberg, FRG).

The different monoclonal antibodies used were usually
applied as undiluted hybridoma culture supernatants.

The secondary antibodies were affinity-purified goat an- /■
tibodies raised against mouse F(ab') 2 , and conjugated to
lissamine rhodamine sulfonyl chloride as previously de-
scribed [5, 20].

Immunohistochemistry

Frozen sections of tissue blocks were cut at about -20° C : V
in a Frigocut 2700 cryostat (Jung-Reichert, FRG). The sec- '.\
tions (4-5 um thick) were placed on clean glass slides, air : j
dried, acetone fixed, and immunolabeled as previously de- : |
scribed [15]. Antibody-stained sections were dehydrated in |
absolute ethanol, mounted in Entelan (Merck, FRG) and : :|
examined using a Zeiss Photomicroscope III equipped for ^
epifluorescence observations with oil-immersion Plan Neo- u
fluar objectives ( x 25/0.8 or x 16/0.5).

ft/*/ eli

jisectedi
||high-s
|-atin c<
I'dimen
I;:f48].
f|
ii*E!ectr

fl^Fresh:
Ifblocki
i|'hyde i
'Ip.postfi
■ence,
fusing;
;voltag

iKesulj

pThehi
l^ytaplasj
lin Fig

p. of not

. 0 - : nonst
|,:|||: 8.60
§p!. ; . grouj:

(pig-'

: regioi
'M 18,18
|p:- canth

fill, mous

serve* |

£11 T
bodie
v d, res
| : with
I and
I exam

Material may b#

Stediby 'cbpynght taw .(Title 17, U.S. Code)

ibody
in all
nalig-
kera-
tides,
: well
2].

kcra-
:sted,
ation

Hides
tiniz-

:* cul-
a not
basal
rrvix,
with
nans.
5 not
cally
-blot
anti-
/ersi-
ss of
M as
n.

;t ey-
loma
iman

*acts
sim-
layer
helia
was
nccr

aally

t an-
d to
de-

0°C
sec-
, air

de-
d in
and

for
Jeo-

HfjSe/ electrophoresis and immunoblotting analyses

liiRelevant regions of 20-um-thick frozen sections were dis-
IltlSected out under microscospic control and extracted with
||v jiigh-salt buffer and detergent [33]. Analysis of the cytoker-
f|ll&tin composition in the sections was carried out using one-
plidimensional gel electrophoresis [27] and immunoblotting
H[48].

m$fc\Ji!ectron microscopy

;

Freshly obtained surgical samples were dissected into small
|||j|>locks (2-3 mm) and immediately fixed in 2% glutaralde-
^p.; fcyde in 0.1 M cacodylate buffer, pH 7.2. The samples were
||; j>ostfixed in 1% 0s0 4 , embedded in Poly bed 8.12 (Polysci-
ISience, USA), cut at the desired orientation, and examined
||; using a Phillips 410 electron microscope at an accelerating
Iff voltage of 80 kV.

Kesults

|| The histological appearance of the various normal and me-
Hftaplastic epithelial elements of the human cervix is shown
"in Fig. 1.

^Cytokeratin expression in the epithelial elements
fof normal human cervix

To establish the pattern of cytokeratin expression as re-
vealed by immunofluorescence labeling, frozen sections of
various regions along the cervix were stained with the six
cytokeratin-specific monoclonal antibodies.

The normal exocervix, throughout its entire length, had
an appearance typical of stratified -squamous epithelium,
with a distinct layer of basal cells and well-ordered supraba-
sal squamous cells (Fig. 1 a). All epithelial layers of the exo-
cercix were intensely labeled with the broadly cross-reacting
K G 8.13 cytokeratin antibody (Fig. 2a) as well as with the
* stratification-specific ' antibody, K s 8.12 (Fig. 2b). Two of
the antibodies used, K M 4.62 and K B 8.37, exclusively la-
beled the basal cell layer (Fig. 2c and e, respectively). It
should be pointed out, however, that in other squamous
epithelial tissues, there are marked differences between the
staining patterns produced by these last two antibodies;
antibody K M 4.62 stains most simple epithelia but does not
label keratinizing squamous epithelia (e.g., epidermis),
while K B 8.37 labels the basel cell layer of keratinizing and
no n keratinizing squamous epithelia but is negative in all
nonsquamous epithelia. Staining of the exocervix with K K
8.60 produced sporadic labeling of individual cells or
groups of cells within the suprabasal layers of the exocervix
(Fig. 2f); the extent of this labeling varied somewhat from
region to region and from sample to sample. Antibody K s
18.18 (reactive with only polypeptide no. 18 did not signifi-
cantly or reproducibly stain any component of the squa-
mous epithelia of the exocervix (Fig. 2d). Occasionally,
faint staining of the basal cells of the exocervix was ob-
served (see Fig. 2d insert).

The normal endocervix was uniformly positive for anti-
bodies Kg 8.13, K M 4.62, and K 8 18.18 (Fig. 3 a, c, and
d, respectively). No labeling of the endocervix was obtained
with the other three antibodies testes, i.e., K s 8.12, K B 8.37
and K K 8.60 (Fig. 3b, e, and f, respectively). Histological
examination of HE-stained sections of the endocervix often

193

revealed the presence of cuboidal cells within the columnar
epithelium. These cells were situated near the basal portion
of the columnar cells and were not exposed at the surface
of the mucosa (Figs. 1 b and 4 a). These cells, identified as
being reserve cells, exhibited the same labeling pattern as
columnar cells with all of the cytokeratin antibodies testes
(Fig. 4a-f).

A high-resolution view of these reserve cells of the endo-
cervix was obtained using transmission electron microsco-
py. Examination of the endocervical mucosa indicated that
these cells were cuboidal cells with electron-lucent cyto-
plasm and a large, round nucleus (Fig. 5a). These cells did
not reach the luminal surface of the endocervix, nor were
they directed attached to the basement membrane (Fig. 5 b,
arrowheads). Examination of a large number of samples
indicated that the reserve cells were trapped' between the
columnar cells, and were attached at their basal aspects
to membrane projections and lamellae of the columnar cells
(Fig. 5b; see Discussion).

Reserve cells were usually sparsely distributed along ma-
jor parts of the endocervix, and the unequivocal identifica-
tion of individual cells was often difficult. However, we
occasionally detected endocervical regions in which various
degrees of reserve-cell hyperplasia were apparent (Fig. 4f,
arrowheads). This manifested itself by a local accumulation
of cuboidal cells in one or a few layers, in which the 'nor-
mal ' columnar cells could be detected at their mucosal as-
pect (Fig. 4a). Staining of hyperplastic reserve cells with
the cytokeratin antibodies revealed positive reactivity with
K c 8.13, K s 18.18, and K M 4.62 (Fig. 4f, arrowheads) as
in the endocervical mucosa (Fig. 3). No labeling of hyper-
plastic reserve cells was observed with antibody K s 8.12
(Fig. 4c) or with antibodies K B 8.37 and K K 8.60 (data
not shown).

Cytokeratin expression in squamous metaplasia
of the human cervix

We distinguished four types of squamous metaplasia. In
the first two metaplasia of the TZM in continuity with
the normal exocervix (type a; Figs. 1 c, e and 6a--e) and me-
taplasia situated within the endocervical canal and its invag-
inations at a distance from the MP (type b; Figs. 1 f and
60, the metaplasia exhibited diminished maturation and
lacked a definite basal layer when compared to normal exo-
cervical squamous epithelium (Fig. 1 c). We also found a
more mature type of metaplasia of the transformation zone,
with distinct, hyperplastic basal cells occupying more than
the usual one layer of cells (type c; Fig. 7). Finally, we iden-
tified metaplasia occurring in glandular elements of an en-
dometrial carcinosarcoma (type d; Fig. 8).

Immunofluorescence labeling of metaplasia of the first
category (type a) with the cytokeratin antibodies resulted
in the metaplastic cells being extensively labeled with K G
8.13 (Fig. 6a), K s 8.12 (Fig. 6c), and K M 4.62 (Fig. 6e, f).
No labeling was obtained with the basal-cell-specific anti-
body, K B 8.37 (data not shown), and individual positive
cells were detected throughout the sections after labeling
with antibody K K 8.60 (Fig. 6d). Staining with antibody
K„ 18.18 produced essentially no labeling of most of the
metaplastic cells (Fig. 6 b), although in some cases, faint
staining of the basal cell layer of the squamous metaplasia
was noticed. Occasionally, strongly labeled residual endo-
cervical epithelial cells were detected at the luminal aspect

Fig. la-f. Light-microscopic appearance of HE-stained sections of normal hyperplastic and metaplastic human cervix The reeions
examined were near to the squamocolumnar junctions (a, c, e) and in the cndc<*rvical canal (b, d/n a Z juncJon between the

^m^s^^s^ss? and . the . simplc epi r, ium ° f i be endocervix tss

area (xzzs). b Endocervix displaying both simple columnar epithelium and a layer of reserve cells (arrowhead- x22S) cSmiammis
metaplasia at the transformation zone (TZM). The arrow indicates a site of previous biopsy ( x60). iR^T^h^rS^^ZZal
in the endocery.cal canal; not.ee the apparently normal epithelium at the top (x 150). e Squamous SS^nSSShS^
the residual columnar cells ( x 240). f Metaplastic plaques (MP) in the endoce^cal glands &rrZe J x ?$ '

' 'M¥teriah^^B^^'6ted :: £y copyright law (Title 1 7, U.S. Code)

gions ^i'iBiJtll-lFig. 2a-f. Immunofluorescence microscopic labeling of the exocervix with monoclonal antibodies, a K G 8.13; b K s 8.12; c K M 4.62;

n the K s 18.18; eK B 8-37 (insert in e shows staining of filaments in the basal cell layer at a higher magnification); f K K 8.60. Note

iction -that antibody K G 8.13 uniformly labeled all of the epithelium, while K s 8.12 stained the suprabasal layers more intensely. K M 4.62

mous ^ift -and K B 8.37 stained only the basal layer, while antibody 8.60 stained individual cells or groups of cells. Antibody 1^ 18.18 was

head) xflifSf! essentially negative, except for occasional faint labeling of the basal cells (insert in d). epithelium; 5, stroma. The arrows in b

icates IP and f point to the basal lamina. Bars, 25 um

«g.3^f .Immunofluorescence microscopic labeling of frozen sections of endocervical simple epithelium with monoclonal antibodies.:
aK ° 8 : , . 3; ,. b , Ks „ * n '\ C L K " 4 62 : d K» 18 - 18 ; e K " 8 -3 7 ; f 8.60. Note the positive reaction of K G 8.13, K M 4.62, and K. 18.18
with epithelial cells, this being in contrast to antibodies K s 8.12, K B 8.37, and K K 8.60 which were negative. S, stroma; E, epithelium;

L, lumen. Bar, 25 \im K

''"Matenafinay te proiected by copyright law (Title 17, U.S. Code)

jpRg. 4a-f. Hematoxylin-eosin staining (a) and immunofluorc
lather normal or hyperplastic reserve cells, b K G 8.13; c
|eells (arrowheads) by antibodies K G 8.13, K % 18.18, and K M

:&V_ ft M A\A nnl 1nlv>l pnithnlial rpllc in th*» pnHfirprviY F. e*n\

1

scence labeling (b-f) with monoclonal antibodies of the endocervix revealing
ft s 8.12; d K s 18.18; e, f K M 4.62. Note the extensive staining of reserve
4.62. The double-arrowhead in f indicates reserve cell hyperplasia. Antibody
. : . : ,:« s »«v .» j ~, - r .thelium; S, stroma; L, lumen. Bar, 25 um

199

i. (The;:
t reach v
ociated

il!:>Fig. 6a-f. Immunofluorescence microscopic labeling of cervical squamous metaplasia using different monoclonal antibodies, a K G 8.13;
||; "b K s 18.18; c K s 8.12; d 8.60; e, f K M 4.62. Note that antibodies K 0 8.13 and K, 8.12 stained the squamous metaplasia uniformly,
t^Avhile K, 18.18 stained only the residual columnar cells of the simple epithelium of the endocervix. Antibody K K 8.60 labeled individual
" [cells or groups of cells in the supra basal layers of the metaplasia, and K M 4.62 uniformly stained all cells of the squamous metaplasia,
|: :both at the transformation zone (e) and in the endocervical canal (f). Note the sharp boundary between the negative suprabasal cells
| : : : : of the exocervix and the metaplastic cells in e. Ex, exocervix; S, stroma; M> metaplasia. Bars, 25

1

? : : M 17; VS. Code)

Fig. 7a-f. Hematoxylin-eosin staining (a) and immunofluorescence microscopy (b-f) of a transformation zone metaplasia exhibiting
a high degree of squamous differentiation, b K c 8.13; c K M 4.62; d K s 8.12; e K B 8.37; f K K 8.60. Note that K G 8.13 and Ki? : " M:
8.12 uniformly stained the squamous metaplasia, while K M 4.62 and K B 8.37 stained predominantly the basal layers. Antibody K K
8.60 stained individual cells or groups of cells within the suprabasal region. 5, stroma. Bars, 25 jim

Fig. 8

||pas the

li : .- onlv 1
H£ S t strc

hibiting :

and K5.;;;
o6y Kg:;;;

i;i Kg. 8a-f. Hematoxylin-cosin (a) and immunofluorescence microscopy (b-f) of a squamous area within an endometrial adenocarcinoma.
II 1>K G 8.13; c K s 8.12; d K, 18.18; e K M 4.62; f K K 8.60. Note that antibodies K G 8.13 and K M 4.62 stained the metaplasia as well
If- as the surrounding adenocarcinoma, while K s 8.12 stained only the metaplasia. K K 8.60 stained individual cells in keratinizing foci
s only in the metaplasia. K s 18.18 stained the metaplasia faintly as compared to its intense labeling of adenocarcinoma cells. Af, metaplasia;
stroma; Ad, adenocarcinoma. Bars, 25 um

202-

CN

00
CO

£0
O

o

CO
00

: ■ ; i

Fig. 9. Immunoblotting analysis of human cervical metaplasia cy-
tokeratins using antibodies K s 8.12 and K K 8.60. Metaplastic re-
gions were microdissected and examined by one-dimensional gel
electrophoresis. The Coomassie-blue (C5)-stained gel contained
polypeptides in the 57- to 59-kilodalton range (corresponding to
polypeptides 5 [35J and 10/11; upper dot) as well as in the ~ 54-kilo-
dalton area (comigrating with polypeptide no. 13; second dot from
top). The two lower dots mark the position of polypeptides nos. 16
(48-kilodaltons) and 19 (40-kilodaltons) which are barely detect-
able by Coomassie-blue staining. The double-arrowhead on the right
indicates the presence of polypeptides nos. 10 and 1 1 in comparable
amounts. The reaction with antibody K s 8.12 shows the major
reactivity of the antibodies with polypeptide no. 13 (upper arrow-
head). The lower band (lower arrowhead) com i grated with polypep-
tide no. 16

Monoclonal Polypeptide
antibody specificity

K c 8 * 13 1.5.6.7.8
(10. 11), 18

K s 8 • 12 13. 16

Ex M En

Ex M g.

Ex M En

4 . 62 19

K B 8 • 37 1

Kg 8 • 60 10, 11

Ex M En

X. 18 • 18 18

\

Fig. 10. Schematic diagram showing the different patterns of stain-
ing produced by the monoclonal antibodies used in this study.
The different regions of the cervix, exocervix (Ex\ metaplasia (A*),
and endocervix (En) are marked, and positive reactivity is indicated
by the shaded areas. The partial shading of basal cells with antibody
K, 18.18 represents the faint occasional labeling obtained with this
antibody. The polypeptide specificities refer to the nomenclature
of Moll etal. [33,40]

meti

§|Immui
tractcc
l| ; :atin p<
lijor gi-
ll'; spond:
detect

In. a;

Ipand a
f£4(M6i
^ analys
;|£ which
!;•: and
fv;: result!
pi'ly rec
I'ix.of no
T body)
II" to a
& nos. U

Discu

of the metaplasia (Fig. 6 b). Results identical to those ob-
tained with TZM metaplasia were also obtained with cases
of MP metaplasia (type b), in which the squamous metapla-
sia was situated at different sites along the cervical canal
and its invaginations (data not shown). The information
obtained from these observations indicated that metaplastic
cells do not express the same combination of cytokeratin
polypeptides as any particular cell type of the normal cervix.
Metaplastic cells were positively labeled with antibody K M
4.62 but were negative or nearly negative for K 8 18.18, un-
like the normal mucosa which was positive for both anti-
bodies. In contrast to the basal layer of the exocervix, the
metaplastic cells were not labeled with antibody K B 8.37.
In being uniformly positively labeled with antibody K s 8.12,
metaplastic cells differed from the columnar and reserve
cells of the endocervix, both of which were not labeled
with this antibody (see Discussion).

Another form of metaplasia (typec) involved a more

mature type of TZM. This metaplasia retained some order ;
of layers, contained prominent basal cells, and was general- ■
ly similar to the neighboring normal exocervix. It could, ;
however, be distinguished from the normal exocervix both
by its anatomical location and by its less ordered strati flea- :
tion (Fig. 7a). fhe staining patterns of this metaplasia with
the battery of cytokeratin antibodies used was different
from that of the 'common 1 forms of metaplasia described
above (types a and b); while all were positively labeled
with antibodies K G 8.13 and K s 8.12 (Fig. 7b, d), typec;
was only partly positive with K M 4.62 (Fig. 6c), and its
basal cells were positively stained by antibodies K B 8.37 : ;
(Fig. 6e) and K, 18.18 (not shown). This is in contrast with •
the common forms of metaplasia which were uniformly pos- j
itive with antibody K M 4.62, and negative with K B 8.37/
Antibody K K 8.60 labeled individual cells throughout the :;|
metaplasia in all three forms (Fig. 7f).

Finally, we examined the transformation of malignant 1

."■ V .iMatesff^^rfffiiK; &i^|j#b@i^#tf Jg^! 'Sdpy ri^ftt .law .0Tttle U.S. Code) . :.;

203

En

En

Dfstam-; :
s study^*;
sia

idicaU$i;
ntibod^p

nclaiurei:;

e order :
;enera|r : :
coul<| •
ix boflM
■atifig^
;ia with - :
iffereoti
scribecj;;;::
label^

typefl
and it$|

LSt wi% : ;
ily pofe:
B 8-371

3Ut tfe||

jgpithelia, rather than normal simple epithelia, into stratified
Isquamous epithelium. Figure 8 a shows a glandular region
||&ithin a carcinosarcoma of the endometrium exhibiting dis-
llinct foci of squamous metaplasia. Immunofluorescence la-
l&eling of this tissue with the cytokeratin-specific antibodies
flfig. 8b-i) produced exactly the same staining pattern ob-
tained in the TZM and MP of the endocervix described
Ipbove.

immunoblotting analysis of cytokeratin polypeptides present
|gj metaplastic cervical tissues

llimmunoblotting analysis of microdissected, high-salt-ex-
Jlfracted tissue sampes revealed the major groups of cytoker-
j&jrtin polypeptides present in this metaplastic tissue. The ma-
imer groups of polypeptides detached were bands corre-
sponding to polypeptide no. 5, which has previously been
l<ietected in this tissue [35], as well as to polypeptide no. 10/
I'll. A second group was found near polypeptide no. 13,
fund a few smaller polypeptides with molecular masses of
lljO-46 kilodaltons were barely detectable. Immunoblotting
^analysis of this sample was carried out in order to determine
which of the polypeptides recognized by antibodies K s 8.12
land K K 8.60 was actually present in the metaplasia. The
flresults (Fig. 9) showed that antibody K s 8.12 predominant-
jlly recognized polypeptide no. 13 and only small amounts
J of no. 16 (both of which react with this particular anti-
|l body). Antibody K K 8.60, on the other hand, was bound
| to a polypeptide doublet corresponding to cytokeratins
|: nos. 10 and 11 (present in essentially equal amounts).

Discussion

. The present study focused on a relatively common type
;|-of 'transdifferentiation' event which occurs in the human
i: cervix, i.e., the development of squamous metaplasia. The
f: major tool used for studying the nature of the metaplastic
|; process was the immunocytochemical and biochemical
\ identification of the cytokeratin polypeptides expressed by
•j normal cells of the cervix and their metaplastic derivatives.
; As pointed out in the Introduction, the expression of cyto-
h keratins in different epithelia has proved to be a most useful
I marker both of the histogenetic origin of cells as well as
f of their state of differentiation [47, 50]. In previous studies
I carried out in several laboratories, the various cytokeratins
h expressed in normal and pathological specimens of the hu-
ll; man female genital tract have been identified [10, 30, 35,
1: 39]. In accordance with the present findings, these studies
I: have indicated the widespread occurrence of the 40-kilodal-
I; ton cytokeratin (no. 19) in the endocervix, metaplastic cells,
i and the basal layer of the exocervix, as well as the presence
; of cornifying foci in normal exocervix and in squamous
; metaplasia.

In the present study, we applied a battery of monoclonal
imtibodies with restricted and defined polypeptide specifici-
ties. Staining of normal and metaplastic cervical tissue with
these antibodies revealed several interesting features relat-
|:i ing to the process of squamous differentiation in general
k and to the formation of squamous cervical metaplasia in
* particular.

A striking property of squamous metaplasia was re-
: vealed by the occurrence of a cytokeratin-polypeptide com-
f bination which is markedly different from that found in
j the epithelial components of the normal cervix. This is sche-

matically illustrated in Fig. 10, which shows the labeling
patterns obtained in the exocervix, in metaplasia, and in
the endocervix using our six monoclonal antibodies. The
marked differences between the metaplastic cells and the
normal cervical components indicate that, regardless of the
nature of the cellular precursor of the metaplasia, the pat-
tern of cytokeratin expression in cells changes during me-
taplastic transformation. Thus, the metaplastic lesions ex-
hibit a largely nonpolar expression of cytokeratins, this be-
ing in contrast to the exocervix; the basal cell layer of the
exocervix was positively labeled with antibodies K B 8.37
and K M 4.62, whereas the metaplasia was uniformly nega-
tive with the former and uniformly positive with the latter
(Fig. 6e). The only suggestion of a limited degree of differ-
ential expression of certain keratins in distinct regions of
the metaplasia was the faint, often barely discernible label-
ing of its basal cell layer with antibody K 8 18.18 (Fig. 6b),
and the sporadic labeling with K K 8.60 (Fig. 6d). Compari-
son of the metaplasia with cells of the normal endocervix
revealed remarkable differences, the most conspicuous of
which was the expression in the metaplasia of cytokeratin
no. 13 and its apparently diminished expression of cytoker-
atin no. 18, which is abundant in normal endocervical mu-
cosa. From these findings, we propose that the formation
of squamous metaplasia represents a new route of differen-
tiation which differs from those detected in the various epi-
thelial elements of the normal cervix. The cells which are
induced to undergo metaplastic squamous differentiation
probably reside in the endocervix. This hypothesis is based
on anatomical considerations and was corroborated by the
results of antibody labeling. Our study of a large number
of cases indicated that metaplastic lesions with similar mor-
phologies and identical cytokeratin patterns may develop
at a distance from the squamocolumnar junction and may
even be detected within adenocarcinomas, thus excluding
the possibility that squamous metaplasia may be formed
by a lateral migration of the exocervix. However, we cannot
at present exclude the possibility that the latter process is
responsible for the formation of the 4 mature' squamous
metaplasia of the transformation zone. This mature form
of metaplasia may develop either by further differentiation
of the 'common' form of metaplasia or by displacement
of the exocervical epithelium.

Findings for human cervical metaplasia in combination
with the results presented here suggest that reserve cells
may be at least bipotent. They may normally terminally
differentiate into columnar, mucous-secreting cells, but
under certain circumstances, they may adopt a stratification
pattern of differentiation and form metaplasia [12, 38]. This
view is corroborated by the finding of apparently intermedi-
ate stages in the development of metaplasia, i.e., reserve
cell hyperplasia (Figs. Id, 40- On the basis of the results
of antibody labeling and immunoblotting following gel-
electrophoresis analyses, it is further proposed that,
throughout the stratification process, new cytokeratins, in-
cluding polypeptides nos. 5 and 13, small amounts of
nos. 16 and 17, and finally, nos. 10 and 11, are gradually
co-expressed. These polypeptides appear in metaplasia
along with three (nos. 7, 8, and 19) of the four cytokeratins
initially present in the endocervical mucosa. The expression
of cytokeratin polypeptide no. 18 decreases during the
course of the metaplastic process. Interestingly, the paired
polypeptides, nos. 5 and 13, are co-expressed in metaplastic
cells; studies in several laboratories have indicated that

Code)

204 '

A

there are cytokeratin polypeptides pairs that are commonly
co-expressed in a differentiation-restricted fashion [50]. The
positive staining of groups of metaplastic cells by antibody
K K 8.60 suggests that, following stratification, another step
towards keratinization may occur, this being manifested
by the appearance of cytokeratins nos. 10 and 11, which
are commonly found in keratinizing squamous epithelia (see
[24, 47, 50]). At least some of these polypeptides (i.e., nos. 5,
7, 8, 17, 18, and 19) have also been detected using two-
dimensional gel electrophoresis [35]. However, the use of
various monoclonal antibodies, particularly K„ 8.12, K K
8.60, and K M 4.62, revealed unequivocally the presence of
the stratification-specific polypeptides nos. 13 and 16, poly-
peptides nos. 10/11, and polypeptide no. 19, respectively.

Another aspect highlighted by our findings involves the
relationships between squamous cervical metaplasia and
neoplastic lesions of the cervix. These will be discussed at
two levels: first, the capacity of squamous metaplasia to
transform into a neoplastic lesion, i.e., squamous cell carci-
noma, and second, the capacity of malignant simple epithe-
lia (i.e., adenocarcinoma) rather than normal simple epithe-
lia to undergo squamous differentiation.

The possibility that squamous metaplasia may consti-
tute a preneoplastic site which is prone to malignant trans-
formation is supported by clinical and histopathological
data [7, 44]. The suggestion of a common origin for squa-
mous metaplasia and neoplasia is corroborated by the pres-
ent findings as well as by the results of recent studies con-
cerning the expression of cytokeratins in squamous cell car-
cinomas and adenocarcinomas (unpublished results). Most
prominent in this respect is the positive labeling of both
lesions with antibody K s 8.12 and their common expression
of the 40-kilodalton (no. 19) polypeptide [33] which is
stained by antibody K M 4.62 [21]. This is also in line with
a previous report of diminished levels of cytokeratin no. 18
in nonkeratinizing squamous cell carcinoma of the cervix
[35]. It is thus concluded that metaplasia and neoplasia
may exhibit a similar pattern of differentiation despite the
marked difference in their proliferative and invasive proper-
ties.

The other aspect, which can only be briefly dealt with
here, concerns the development of squamous metaplasia
within adenocarcinoma. The existence of adenocarcinomas
with regions exhibiting squamous differentiation has been
well established by conventional histopathology [26]. The
data presented in our study support the view that such
metaplastic cells express polypeptides largely similar to
those of 'con ventionaP squamous metaplasias derived from
normal, nonneoplastic epithelium. This is mainly shown
by the persistent expression of cytokeratin no. 19 and the
positive labeling with antibody K s 8.12. We observed vari-
able degrees of labeling with antibodies to polypeptide
no. 18, although we still do not know whether this should
be attributed to the fact that the tissue of origin studied
here was derived from the endometrium rather than the
endocervix or to the fact that is was malignant. This aspect
is now under investigation. It might, however, be concluded
that virtually the same changes in cytokeratin expression
occur during metaplastic transformation regardless of
whether the epithelium of origin is normal or neoplastic.

In conclusion, the present study shed light on basic pro-
cesses of squamous differentiation from its early stages,
characterized by the hyperplastic growth of reserve cells
through stratification, to the development of focal keratin-

izing centers, as well as on the molecular relationships be-
tween squamous metaplasia and neoplastic transformation
of the cervix.

Acknowledgements. We would like to thank W.W. Frankc (German
Cancer Research Center, Heidelberg) for the generous gift of
monoclonal antibody it, 18.18. This study was supported in pan
by a grant from the NCRD-BMFT Israeli-German Joint Project
Program. J

References

1. Altmannsberger M, Osborn M, Schauer A, Weber K (1981)
Antibodies to different intermediate filament proteins. Cell
type-specific markers on paraffin-embedded human tissue*
Lab Invest 45:427-434

2. Anderton BH (1981) Intermediate filaments: A family of ho*
mologous structure. J Muscle Res Cell Motil 2:141-166

3. Auerbach O, Gere JB, Forman JB (1957) Changes in the bron-
chial epithelium in relation to smoking and cancer or the lune
N Engl J Med 256:97-104 g *

4. Battifora H, Sun T-T, Bahu RM, Rao S (1980) The use of
antikeratin antiserum as a diagnostic tool: Thymoma versus
lymphoma. Hum Pathol 1 1 : 635-641

5. Brandtzaeg P (1973) Conjugates of immunoglobulin G with
different fluorophores. I. Characterization by anionic exchange
chromatography. Scand J Immunol 2:273-290

6. Castano-Almendral A, Muller H, Naujoks H, Castano-AImen-
dral JL (1973) Topographical and histological localization of
dysplasias, carcinomata in situ, microinvasions and microcar-
cinomata. Gynecol Oncol 1 : 320-329

7. Christopherson WM, Healon N, Gray LA Sr (1979) Non-inva-
sive precursor lesions of adenocarcinoma and mixed adeno-
squamous carcinoma of the cervix uteri. Cancer 44 : 975-983

8. Coppleson M (1980) The natural history of premalignanl le-
sions of the genital tract. A colposcopist's view. Exerpta Med
Int Congr Ser 51 2 : 304-307

9. Correa P, Cuello C, Duque E (3970) Carcinoma and intestinal
metaplasia of the stomach in Colombian migrants. J Natl Can
Inst 44:297-306

10. Dixon IS, Margaret AS (1984) Immunofluorescent studies of
human cervical epithelia in vivo and in vitro using antibodies.:
against specific keratin components. Mol Biol Med 2 * 37- 51

11. Eversole LR (1978) Clinical outline or oral pathology: Diagno-
sis and treatment. Lea and Febiger, Philadelphia, pp 56-57

12. Ferenczy A (1982) Anatomy and histology of the cervix and ,
cervical intraepithelial neoplasia. In: Blaustcin A (ed) Patholo- '
gy of the female genital tract, 2nd edn. Springer, Berlin Heidel-
berg New York, pp 1 26-1 32

13. Ferenczy A, Richert RM (1984) Female reproductive system: -
Dynamics of scan and transmission electron microscopy J Wi- ■'.
ley and Sons, New York, pp 66-68

14. Fluhmann CF (1954) Comparative studies of squamous meta- :=
plasia of the cervix, uteri and endometrium. Am J Obstet Gyne-
col 68: 1447-1463

15. Franke WW, Appelhans B, Schmid E, Frcudenstein C, Osborn
M, Weber K (1979) Identification and characterization of epi-
thelial cells in mammalian tissues by immunofluorescence ^
microscopy using antibody to prekeratin. Differentiation
15:7-25

16. Franke WW, Schiller DL. Moll R, Winter S, Schmid E, Engel- /
brecht I (1981) Diversity of cylokeratins-differentiation specific "'■
expression of cytokeratin polypeptides in epithelial cells and
tissues. J Mol Biol 153:933-959

17. Franke WW, Schmid E, Schiller DL. Winter S, Jarasch ED, %
Moll R, Denk H, Jackson BW, Illmensee K (1981) Differentia- si
tion-related patterns of expression of proteins or intermediate- -i;
sized filaments in tissues and cultured cells. Cold Spring Harbor
Symp Quant Biol 46:432-453

18. Fuchs E (1983) Evolution and complexity of the genes encoding

the ki

|li9. Gabbi
muno
huma
ogist.
||0. Gcige
1 cappii
|l.Gigi-l
twecn
I CelU
!&2. Gigi
Schill
deten
cross*
|p3. Hanu
II cyt
domaj
I&4. Huszi
(1986
in tht

1129.

ifc 30.

31

asscn

Laza:

grate

Laza;

ogen<

Rev

Loni

and 1

Int J;

Luge

108:

Mini

205

ips be-
m ation :

je rman
gift of
in part ;

(1981)
■s. Cell
tissues; ■

of hoB :

use of ;
versus i

j with. ;:
:hangei;

Vlmen v
ion of

:roear;; :

i-inva- : :
ideno* ! V

983

int le* ! .
i Med -

islinal v
il Can"*"

stem:
J Wi-

the keratins of human epidermal cells. J Invest Dermatol
81 :141s-144s

Gabbiani G, Kapanci Y, Barazzone P, Franke WW (1981) Im-
munochemical identification of intermediate-sized filaments in
human neoplastic cells : A diagnostic aid for the surgical pathol-
ogist. Am J Pathol 104:206-216
1§0. Geiger B, Singer SJ (1979) Participation of a-actinin in the
W capping of membrane components. Cell 16:21 3-222
|||. Gigi-Leitner O, Geiger B (1986) Antigenic interrelationship be-
tween 40 kilodalton cytokeratin polypeptide and desmoplakins.
.... Cell Motil and the Cytoskeleton (in press)
|2 Gigi O, Geiger B, Eshhar Z, Moll R, Schmid E, Winter S,
|fe Schiller DL, Franke WW (1982) Detection of a cytokeratin

determinant common to diverse epithelial cells by a broadly
cross-reacting monoclonal antibody. EM BO J 1 : 1429-1437
Hanukoglu 1, Fuchs WV (1983) The cDNA sequence of a type-
II cytoskeletal keratin reveals constant and variable structural
domains among keratins. Cell 33:915-924
LHuszar M, Gigi-Leitner O, Moll R, Franke WW, Geiger B
(1986) Polypeptide-specific monoclonal cytokeratin antibodies
in the differential diagnosis of squamous carcinomas and ade-
nocarcinomas. Differentiation (in press)
il25 Koss L (1975) Tumors of the urinary bladder. Atlas of tumor
|; : ' pathology, second series, fascicle 11. Armed Forces Institute
M of Pathology, Washington, pp 103

|j> 26. Kurman RY, Norris HI (1982) Endometrial neoplasia, hyper-
plasia and carcinoma. In: Blaustein A (ed) Pathology or the
;W female genital tract. Springer, Berlin Heidelberg New York,
"I pp 331-333

i;27. Laemmli UK (1970) Cleavage of structural proteins during the
Jf assembly of the head of bacteriophage T4. Nature 227 : 680-^85
§1*28. Lazarides E (1980) Intermediate filaments as mechanical intc-
grators of cellular space. Nature 283 : 249-255
f 29. Lazarides E (1982) Intermediate filaments: A chemically heter-
ogeneous, developmenially regulated class of proteins. Annu
Rev Biochcm 51 : 219-250
11: 30. Loning TH, Kohler CH, Casclitz J, Stegncr HE (1983) Keratin
and tissue polypeptide antigen profiles of the cervical mucosa.
Int J Gynecol Pathol 2 : 1 05-1 1 2
31. Lugo M, Putong PB (1984) Metaplasia. Arch Pathol Lab Med
108:185-189

| 32. Ming SC, Goldman H, Freiman DG (1967) Intestinal metapla-
* :: sia and histogenesis of carcinoma in human stomach: Light
and electron microscopic study. Cancer 20: 1418-1429

33. Moll R, Franke WW, Schiller DL, Geiger B, Krcpler R (1982)
The catalog of human cy tokeratins : Patterns of expression in
normal epithelia, tumors, and cultured cells. Cell 31 : 1 1-24

34. Moll R, Krepler R, Franke WW (1983) Complex cytokeratin
polypeptide patterns observed in certain human carcinomas.
Differentiation 23:256-269

35. Moll R, Levy R, Czernobilsky B, Hohlweg-Majert P, Dallen-
bach-Hellweg D, Franke WW (1983) Cytokeratins of normal
epithelia and some neoplasms of the female genital tract. Lab
Invest 49:599 -610

36. Osborn M (1983) intermediate filaments as histologic markers:
An overview. J Invest Dermatol 81 : 104s-109s

37. Osborn M, Altmannsberger M, Debus E, Weber K (1984) Con-
ventional and monoclonal antibodies to intermediate filament

proteins in human tumor diagnosis. In : Cancer cells, the trans-
formed phenotypc, vol 1. Cold Spring Harbor Laboratory,
New York, pp 191-200

38. Philipp E (1975) Elektronen-mikroskopische Untersuchungen
fiber die sogenannten Reservzellen und Zylinderepithel des
menschlichen Cervix Uteri. Arch Gynecol 218:295-311

39. Puts JOJG, Moesker O, Kenemans P, Vooijs PG, Ramaekers
FCS (1985) Expression of cytokeratins in neoplastic epithelial
lesions of the uterine cervix. Int J Gynecol Pathol (in press)

40. Quinlan RA, Schiller DL, Hartzfeld M, Achtstatter T, Moll
R, Jorcano JL, Magin TM, Franke WW (1985) Patterns of
expression and organization of cytokeratin intermediate fila-
ments. In: Wang E, Fischman D, Licm RKH, Sun T-T (eds)
Intermediate filaments. Ann NY Acad Sci 455:282-306

41. Ramaekers FCS, Puts JJG, Kant A, Moesker O, Jap PHK,
Vooijs GP (1981) Use of antibodies to intermediate filaments
in the characterization of human tumors. Cold Spring Harbor
Symp Quant Biol 46:331-339

42. Ramaekers FCS, Huysmans A, Moesker O, Kant A, Jap P,
Herman C, Vooijs P (1983) Monoclonal antibody to keratin
filaments, specific for glandular epithelia and their tumors. Lab
Invest 49:353-359

43. Reagan JW, Fu YS (1983) The uterine cervix. In: Silverberg
SG (ed) Principles and practice of surgical pathology, vol 2.
J Wiley and Sons, New York, pp 1223

44. Richart RM (1973) Cervical intraepithelial neoplasia. In: Som-
mers SC (ed) Pathology annual. Appleton-Ccntury-Croft, New
York, pp 301-328

45. Schiller DL, Franke WW, Geiger B (1982) A subfamily of rela-
tively large and basic cytokeratin polypeptides as defined by
peptide mapping is represented by one or several polypeptides
in epithelial cells. EM BO J 1 :761 769

46. Stedman's Medical Dictionary (1982) 24th edn. Williams and
Wilkins, Baltimore, p 864

47. Sun T-T, Eichner R, Schermer A, Cooper D, Nelson WG,
Weiss RA (1984) Classification, expression and possible mecha-
nisms of evolution of mammalian epithelia keratins: A unifying
model. In: Levine A, Topp W, Vande Woude GF, Watson
JD (eds) Cancer cells, the transformed phenotype, vol 1. Cold
Spring Harbor Laboratory, New York, pp 169-176

48. Towbin H, Staehelin T, Gordon J (1979) Electrophonetic
transfer of proteins from poiyacrylamide gels to nitrocellulose
sheets : Procedure and some applications. Proc Natl Acad Sci
USA 76:4350-4354

49. Trump BF, McDowell EM, Glavin F (1978) The respiratory
epithelium. III. Histogenesis of epidermoid metaplasia and car-
cinoma in situ in the human. J Natl Can Inst 61 : 565-575

50. Tseng SCG, Jarvinen MJ, Nelson WG, Huang JW, Woodcock-
Mitchell J, Sun T-T (1982) Correlation of specific keratins with
different types of epithelial differentiation: Monoclonal anti^
body studies. Cell 30:361-372

51. Vogel AM, Gown AM (1984) Monoclonal antibodies to inter-
mediate filament proteins : Use in diagnostic surgical pathology.
In: Shay J (ed) Cell and muscle motility. Plenum, New York
London, pp 397^02

Received March 1986 / Accepted in revised form April 19, 1986

ngel- .
ecific
and :

' = . ■*: .: .' :'1v1aferfa?^ U; US. Code)